Learning from Consensus and Disagreement: Unsupervised On-Policy Self-Distillation with Minority-Trajectory Contrast
Jiaxin Guo, Yanwei Yue, Xuanbo Fan, Chunyu Yang, Yan Zhang
cs.LG, cs.AI
Submitted: 2026-08-09
Updated: 2026-08-11
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- KTO: Model Alignment as Prospect Theoretic Optimization
- Revisiting On-Policy Distillation: Empirical Failure Modes and Simple Fixes
- OpenThoughts: Data Recipes for Reasoning Models
- Distilling the Knowledge in a Neural Network
- Reinforcement Learning via Self-Distillation
- Asymmetric On-Policy Distillation: Bridging Exploitation and Imitation at the Token Level
- UniSD: Towards a Unified Self-Distillation Framework for Large Language Models
- DistiLLM: Towards Streamlined Distillation for Large Language Models
- Training language models to follow instructions with human feedback
- Privileged Information Distillation for Language Models
- Direct Preference Optimization: Your Language Model is Secretly a Reward Model
- DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
- Anti-Self-Distillation for Reasoning RL via Pointwise Mutual Information
- From Generic Correlation to Input-Specific Credit in On-Policy Self Distillation
- GATES: Self-Distillation under Privileged Context with Consensus Gating
- PACED: Distillation and On-Policy Self-Distillation at the Frontier of Student Competence
- TIP: Token Importance in On-Policy Distillation
- Qwen3 Technical Report
- Self-Distilled RLVR
- On-Policy Context Distillation for Language Models
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks